Fuzzy Model Development for a Continuous Dynamic Black-Box System

The paper describes a method of a black-box fuzzy model design based only on the system measured input/output quantity without necessity of any preliminary knowledge about the modeled system order and without any access to system state variables. The method is based on the use of continuous linear dynamic filters of the first order, which in practice are easily feasible and are always stable. The proposed methodology is applied in the fuzzy model design for a three-phase induction motor which represents a higher order nonlinear system. Simulation measurements in MATLAB confirmed the accuracy, effectiveness and good dynamic properties of the proposed fuzzy model and also its applicability to continuous dynamic systems with as little previous knowledge as possible. The aim of the article is to show that the automated fuzzy modeling techniques can be used also to obtain an accurate fuzzy model of a low complexity based on a set of series of measurements.